A Competition Perspective on Physician Non-compete Agreements

Abstract

Physician non-compete agreements may have significant competitive implications, and effects on both providers and patients, but they are treated variously under the law on a state-by-state basis. Reviewing the relevant law and the economic literature cannot identify with confidence the net effects of such agreements on either physicians or health care delivery with any generality. In addition to identifying future research projects to inform policy, it is argued that the antitrust “rule of reason” provides a useful and established framework with which to evaluate such agreements in specific health care markets and, potentially, to address those agreements most likely to do significant damage to health care competition and consumers.

• What do we already know about this topic?

• Relatively little is known about the impact of physician non-compete terms, although 2 recent empirical papers seek to shed some light on the topic.

• How does your research contribute to the field?

• This paper examines physician non-competes in terms of competition policy. That suggests market-specific evaluation under a framework akin to the antitrust “rule of reason,” as well as further topics for research.

• What are your research’s implications toward theory, practice, or policy?

• First, antitrust enforcement has the potential to address cost and access problems associated with physician non-competes where they are most likely to be substantial: specific physician services markets with limited competition. Second, antitrust cases may be viable independent of federal regulatory proposals. Third, given renewed enforcement focus on non-competes, institutional providers should consider the potential for antitrust liability when contracting with physicians and other practitioners.

Introduction and Approach

Non-compete terms (NCTs)—sometimes called “covenants not to compete” or “non-competes”—are much in the news; and recent years have seen both state law reform and a proposal by the Federal Trade Commission (FTC) to ban almost all NCTs. ¹ Provider NCTs provide a unique locus of policy interest, based on familiar health care policy concerns about cost, quality, and access. ² Physician NCTs restrict physician mobility and can impede competition for physician services. While physician services are by no means the largest component of health care spending, they are a substantial part, and impediments to the efficient flow of physician services are an important policy concern.

NCTs have been both permitted and regulated under the common law and state statutory law, based on a recognition that NCTs can implicate both legitimate business interests and risks to competition.^{3 -5} This paper considers physician NCTs through the lens of antitrust law and competition policy. While competition concerns need not exhaust policy considerations, competition policy and antitrust law—at least as established in the U.S.—offer a framework on which to analyze the impact of commercial conduct on competition and, specifically, on competitive effects that harm consumers; namely, in health care services markets, patients, as well as third-party payers. Key elements of the framework are the tools of industrial organization economics and, generally, a focus on the likely effects of commercial conduct. Competition policy more broadly applies the tools and goals of antitrust beyond the scope of commercial conduct to inform public policy.

Section II, below, provides legal and economic background on NCTs. Section III identifies significant research challenges and provides a critical review of the limited empirical literature on physician NCTs. Section IV identifies avenues of research and a way forward independent of general regulations: first, market-specific inquiry under the antitrust “rule of reason” ought to be the primary means of addressing anticompetitive physician NCTs; second, certain physician NCTs in highly concentrated services markets can be subject to heightened scrutiny. Consistent with AMA recommendations, special concerns may be implicated for trainee NCTs, given the constraints the Match imposes on competition for and among residents and fellows.

Legal and Economic Background on NCTs

NCTs and the Law

NCTs are terms in employment contracts limiting the ability of an employee to work for, or establish, a competing firm after separation from the signatory firm.^{3 -5} Historically, U.S. NCTs have been governed by state statutory and common law. NCTs can vary along several dimensions, including duration, geographic scope, income level, occupation, and triggering conditions.^{3 -9} And the conditions under which states permit firms to obtain court orders enforcing NCTs also vary along several dimensions: several states broadly prohibit enforcement by declaring NCTs void; other states limit enforcement for some category of employees, such as low-wage earners, health care practitioners, or tech workers; and most states permit enforcement more widely, typically subject to some sort of “reasonableness” test.^3,5,6,10

Federal NCT regulations have been proposed, if not adopted or applied. Close analysis of the proposals is beyond the scope of this paper, but they are discussed elsewhere at length,^{1,7 -9} and merit mention here. The FTC’s proposed rule “would provide that it is an unfair method of competition . . . for an employer to enter into or attempt to enter into a non-compete clause with a worker; maintain with a worker a non-compete clause; or, under certain circumstances, to represent to a worker that the worker is subject to a non-compete clause.”^{1(p. 3482)} That is, whereas many states restrict civil suits, the FTC proposes to ban NCTs outright, albeit subject to a limited exception with the sale of a business. Independently, the General Counsel of the National Labor Relations Board issued a memo “setting forth her view that the proffer, maintenance, and enforcement of non-compete provisions in employment contracts and severance agreements violates the National Labor Relations Act. . . .” ¹¹

Here, note simply that NLRB pronouncement has not been implemented. Doubtless, it will be challenged in court should the NLRB seek to enforce it, partly due to its broad sweep, and partly because it seems facially dubious. The FTC proposal rests on a claim of general competition rulemaking authority under Section 6 of the FTC Act. That claim is not baseless, but it too has been questioned by many legal scholars,^{7 -9,12} and there are reasons to think that the FTC’s broad proposal would also fail in the courts. For present purposes, note too that FTC enforcement authority against not-for-profit providers is limited.^{1(p. 3510)}

State regulation is varied but well established. For example, the core of California NCT law is a provision in the state’s Business and Profession’s Code from 1941 ¹³ that has both statutory and case law roots dating to the 19th Century. ³ California is 1 of 4 states with general—albeit not universal—bars to NCT enforcement. Those limits are sufficiently broad to apply to most physician NCTs. At the other end of the spectrum are states generally permissive of NCTs, if still subject to some sort of “reasonability” test, which may be more-or-less permissive depending on the state.

Fifteen additional states and the District of Columbia restrict enforcement of at least some physician NCTs.^2,6,14 The details vary considerably. For example, Connecticut, Massachusetts, New Hampshire, Rhode Island, and South Dakota stipulate that physician NCTs are void and unenforceable, subject to a limited exception in the case of the sale of a business—perhaps not so limited for physicians with ownership interests in group practices. D.C. restricts NCTs for any person who is not a “highly compensated employee”—in effect, barring NCT enforcement against physicians who earn less than $250,000 per year (in 2022 dollars). Iowa law does not address physician NCTs generally, but bars enforcement against any “mental health professional,” expressly including psychiatrists.

Several states—Arkansas, New Mexico, South Dakota, and Tennessee—limit the application of NCTs to physicians among other health care professionals, with South Dakota specifying 28 categories of health care practitioners for whom NCTs are “voidable,” except with the sale of a practice. And several states—Indiana, Tennessee, and Texas—permit the enforcement of physician NCTs, but only subject to numerous conditions besides the typical “reasonability” test. In Indiana and Texas, those conditions include provisions to facilitate the continuation of established practitioner-patient relationships.

Whereas Arkansas, New Mexico, South Dakota, and Tennessee significantly constrain default permission to enforce physician NCTs, Colorado and Delaware law appear to impugn enforcement broadly, but subject to a significant qualification. Specifically, Colorado and Delaware stipulate that physician NCTs may provide that the physician is liable for damages, including “damages related to competition.” These last exceptions seem especially vitiating from a competition perspective. That is, “damages” may comprise not only an employer’s lost investments in, for example, training and business development, but an employer’s anticipated loss of economic rents: profits due to provider market power that may be diminished should a physician departure improve competition, to the benefit of patients and other payers.

High-Level Economic Observations

The effects of NCTs on both labor markets and downstream products and service markets are theoretically ambiguous. NCTs can solve or mitigate hold-up problems in research and development and, specifically, in employee training and the sharing of proprietary information with employees.^{3,7,8,15 -17} On the other hand, NCTs can suppress employee mobility, thereby restricting competition for labor in ways that can also impede downstream competition.^{10,18 -20} Empirical research—however challenging—is critical to resolving the theoretical ambiguity, but the empirical literature, while growing, remains limited. Relatively little is known about the impact of physician NCTs specifically, although 2 empirical papers seek to fill the gap.^18,21 While each represents a useful contribution, 2 papers are not a body of literature, much less a settled one: they do not give a comprehensive picture of the impact of physician NCTs on either physician labor markets or health care services markets; and, as discussed below, are subject to significant data and methodological limitations, which constrains their policy implications.

NCTs restrict workers’ ability to compete for jobs; they are, as such, restrictions on supply in diverse labor markets. That is, NCTs appear, facially, “vertical” agreements restraining labor market competition, at least at the margin—if not typically agreements among competitors. Those restrictions are not merely nominal—they are enforceable in most states;^3,4,6,14 they have been enforced in court^{22 -24}; and actual or threatened suits may have chilling effects beyond a specific case or controversy.^{17(pp. 505, 516),25(p. 25)} It is not surprising that they raise efficiency concerns, among others.

Potential harms include reduced worker mobility,^18,20 reduced wages or total compensation,^18,20,26,27 reduced employee bargaining power, and, in downstream services markets, increased cost of entry or expansion, given a restricted supply of skilled or experienced labor, and given that employees bound by NCTs may be unable to establish their own competing firms. ¹⁰

While those concerns are significant, they identify potential problems: their presence and magnitude, and the question of their net effects, appear context dependent. NCTs are not necessarily inefficient, anticompetitive, or harmful to either labor or consumers.^3,7,8,15,28 As noted above, NCTs can solve a range of potential hold-up problems in labor contracting.^7,15,16 For example, because experienced labor is alienable, firms may tend to under-invest in employee training, and may share inefficiently low levels of private information—such as trade secrets or client lists—if unable to recoup their investments before renegotiating compensation and, not incidentally, to limit their competitors’ ability to free-ride on those investments.^3,7,15,16 Employees may, in turn, see reduced wage growth, to the extent that their own investments in education, training, or information do not offset reductions in firm investment. Correspondingly, several studies indicate that NCTs are associated with greater firm investments in employee training.^4,16,29

Several studies suggest that NCTs (or greater “enforceability” of NCTs) tend to diminish worker compensation, at least on average,^16,20,26,27 but results are mixed: one study suggests that the timing of employee notice affects the direction of wage outcomes, ¹⁰ and several studies find NCTs linked to higher compensation for specific categories of professionals,^18,28,30 including physicians. ¹⁸ And while NCTs can reduce the supply of available labor, they can also reduce (or increase) search and training costs by reducing turnover—and those benefits may be shared, at least to some extent, with employees.^15,17

Diverse policy concerns have been raised about physician NCTs, at least in some contexts.^2,31 While these adopt no specific empirical or normative framework, mapping general NCT issues in the literature onto physicians is, in abstract, relatively straightforward. Physician NCTs restrict the ability of physicians to change employment or establish a new practice proximate to their prior practice for some specified length of time. Hence, they restrict a physician’s ability to seek higher compensation or preferred working conditions and, thereby, restrict that physician’s entry into a local labor market. Such restrictions might be bargained away, but not without cost, and they can interact with other restrictions on expansion or entry, such as state Certificate of Need laws.^32,33 At the same time, established rationales on behalf of NCTs, at least in some circumstances, may equally apply.^2,31

Competitive effects may be especially pronounced in highly concentrated provider markets—more so with significant regulatory barriers to entry. Similarly, trainee NCTs might be seen against the backdrop of a statutory antitrust exemption for graduate medical residency matching programs, including the National Resident Matching Program.^34,35 The statute stipulates that these programs are “highly efficient” and “procompetitive.” ³⁵ While the Match does confer a measure of search efficiency for both residents and training institutions, it also constrains competition for and among residents. The statutory declaration notwithstanding, it is unknown whether the Match is procompetitive on net, and adding NCTs could further undermine physician labor market competition.

Still, general concerns about the mitigation of hold-up problems may apply to medical practice specifically, given investments in training, practice development, and the sharing of proprietary information and information systems that are potentially productive and procompetitive. As with NCTs generally, tradeoffs between potential competitive benefits and potential competitive harms may vary across markets; and—pace the FTC’s regulatory ambitions—require context-specific investigation in any case.

The Empirical Record and Empirical Challenges

Physician Studies

Empirical study is critical to NCT policy making, but the empirical record is mixed, and hardly comprehensive. While there is a growing body of economic literature on NCTs, only 2 papers study physician NCTs specifically. For one, Lavetti et al ¹⁸ conducted a survey of primary care physicians in 5 states, gathering input into panel data on both NCT usage and various labor market outcomes, such as earnings. Those data were analyzed with and without the findings on the relative “enforceability” ⁶ of state laws. The results suggest that—at least on average—NCTs are associated with higher physician compensation, estimating “that [NCTs] increase the annual rate of earnings growth by an average of 8 percentage points in each of the first 4 years of a job, with a cumulative effect of 35 percentage points after 10 years on the job.” ¹⁸ Analyzing wage growth in a model using variation in enforceability, the estimates are 89% wage gain over 10 years with NCTs but 36% without. They also find a higher incidence of patient referrals associated with NCTs, which may imply allocative and search efficiencies, and potentially patient benefits.

A related study provides the only empirical evidence—for any occupation—linking NCTs and downstream prices cited in the FTC’s literature review. Hausman and Lavetti ²¹ find that “changes in [NCT] enforcement can provide differential incentives for growth at the establishment versus firm levels.” ²¹ They observe, specifically, 1.7 to 2.1 percent differences in negotiated prices associated with a 10% change in “enforceability” and a corresponding 100-point changes in firm-level Herfindahl-Hirschman Index (HHI)^{21(p. 260)} (which sums the squares of each participant’s market share). The results are similar in scale, but directionally inverse, for changes at the facility level, which may implicate trade-offs between the efficiency gains of larger establishments and the improved negotiating position associated with bargaining as a larger organization.” ²¹ They also observe that the price effects appear to flatten after about 2 years. ²¹

The results are striking, but there are reasons to question the estimates, which suggest significant near-term price effects, on average, from relatively subtle policy and market changes. To put the 100-point HHI change in context, under the FTC/DOJ 2010 Horizontal Merger Guidelines, markets with HHI below 1500 were considered “unconcentrated,” and mergers resulting in unconcentrated markets were considered “unlikely to have adverse competitive effects and ordinarily require no further analysis,” as were mergers involving small changes in concentration.^{36(p. 18)} Although new guidelines released by the agencies revise those thresholds, the empirical basis of the revision is unclear; and they do not, in any event, suggest that small changes in concentration raise competitive concerns without highly concentrated markets. ³⁷ Consider, for example, a market where 10 firms provide general pediatric services. For simplicity, assume that each firm has a 10% market share. In that case, the HHI is 1,000 (10(10²)). Suppose, further, that 2 practices merge, with the 8 non-merging firms each retaining its 10% market share. Then the HHI would be 1,200 (8(10²) +20²). That acquisition would yield double the change that is supposed to be robustly associated with significant price increases, not just in a specific market, but on average.

While small changes in HHI can be competitively significant in threshold cases, perhaps in the presence of high baseline concentration, they do not tend to be. No doubt there are physicians on the margin, ready to switch jobs locally pending release from an NCT, but significant provider reorganization, with price implications, within 2 years, based on subtle changes in “enforceability,” seems unlikely—all the more so given, for example, cycles of health plan reimbursement negotiations, varying lags to effective dates for legislation, variation on whether enforceability changes apply, retroactively, to agreements already in place (many do not), and regulatory—on top of financial—barriers to establishing a new practice. Unlikely results may, of course, prove accurate, but they implicate further inquiry; and they need to be considered against both data and methodological limitations to any single inquiry and general challenges facing the larger body of literature.

Empirical Challenges

The economic literature on NCTs, while substantial, is relatively new; and because available evidence on the use of NCTs is limited, NCTs are commonly studied indirectly, via changes in their “enforceability” under state law. Even that evidence is far from comprehensive or definitive. Space does not permit a thorough critical synthesis of the literature. Still, note that leading contributors to the positive literature^4,17 have been among those raising significant concerns about the available data and methodological limitations.^{3,4,7,8,15,17} While all studies are subject to limitations, significant issues running across an emerging body of literature call into question the extent to which available findings justify policy reform.

Bishara and Starr ¹⁷ observed significant data and methodological limitations across much of the available literature. For example, limited objective data prompt over-reliance on survey data (of varying size and quality) and impede tying policy changes—and subsequent economic indicators—to those workers who are subject to NCTs. Identification issues abound. And many studies cannot properly be said to evidence causal design.

For competition purposes, we might add the question whether average state-wide effects are dominated by specific geographic markets where key employers enjoy outsize market power. ⁷ A working paper from the FTC’s Bureau of Economics raises further concerns: the “more credible empirical studies tend to be narrow in scope, focusing on a limited number of specific occupations . . . or potentially idiosyncratic policy changes with uncertain and hard-to-quantify generalizability.” ¹⁵ These challenges, together with conspicuous lacunae in the literature, constrain our ability to generalize from available findings.

An additional problem is raised by the idiosyncratic policy changes, ¹⁵ identification problems,^15,17 and dearth—at least until recently—of exogenous variation. ¹⁷ Most studies tie average changes for, for example, some tranche of the income distribution, to diverse policy changes. There may be some intuitive appeal to this approach for several reasons. First, there is survey evidence on the incidence of NCT usage within and across states, but little to identify the individuals bound by NCTs, so it is difficult to study the impact of NCTs directly. Second, evidence on the effects of policy reforms should bear directly on the question what policies to adopt. Third, most pertinent state laws address a private party’s ability to enforce an NCT in a civil suit; and at a high level of abstraction, we might have an intuitive sense that some laws are more (or less) restrictive than others, and some jurisdictions more (or less) “plaintiff friendly.” For example, California courts, applying California Business and Professions Code Section 16600, commonly repudiate plaintiffs’ NCT claims. Hence, we might consider California as a “low enforceability” state.

Development of a clear and reliable policy metric is another matter. To deal with myriad legislative and judicial changes in potentially relevant bodies of law,^3,7,17,19 Bishara ¹⁹ developed a measurement scheme that has, in varied implementations, been applied across most of the empirical literature. The scheme depends on several layers of subjective judgments, beginning with Malsberger, ³⁸ a periodic 50-state review of NCT laws, with accompanying questions to structure reviews of NCT laws. Based on 12 criteria of “enforceability” behind those questions, Bishara and Starr examined individual changes in state statutory and (chiefly) decisional law and applied “seven [derivative] questions because they directly address the legal issues relevant to measuring a given jurisdiction’s intensity of noncompete enforcement.” These range from, for example, whether there is a statute of general application to NCT enforcement to the question how an employer’s “protectable interest” is defined. Each state law “event” received a score in response to each question and an aggregate score (a weighted sum of its individual scores). ¹⁷

There are reasons to wonder about the scoring scheme, which has been implemented differently by different researchers. While Malsberger’s questions represent the informed judgment of an experienced attorney, they are subjective judgments that were not developed to serve any specific measurement goals or research projects. Individual legal changes are scored, and weights assigned, as further matters of judgment by non-lawyers. Given the diversity and subtlety of legal changes, even experts might be unable to estimate the various spillover consequences of any given judicial decision against some background legal environment that is not independently assessed. The language in which a decision is framed may be a poor signal of any specific consequences. ³ Fix the judicial language and it can be hard to predict large, small, or infra-marginal effects, depending on the variable of interest. ⁷

Moreover, the studies do not specify the variable of interest. It seems clear that any number of factors or end points might matter: the incidence of litigation, the cost (average, median, or modal) of litigating a case to its conclusion; the ratio of plaintiff to defendant success in litigated cases; the number or frequency of NCT complaints, and, for complaints filed, settled or, perhaps, surviving, for example, motions to dismiss for failure to state a claim, motions for summary judgment, etc. We might also be interested in the way the law affects not just the incidence of NCTs, but the distribution and terms of NCTs. These are all related factors, but hardly interchangeable. A priori, there is no obvious set or function of them that is best for all (or any specific) research project.

The FTC suggests that while “studies have defined enforceability . . . in slightly different ways, each uses enforceability as a proxy for the chance that a given noncompete clause will be enforced.”^{1(p. 3486)} Perhaps, but the studies do not appear to say so expressly, and it is not clear that anyone ever investigated, empirically, the link between “enforceability” scores and legal indicators for which they might be proxy. This is not a Western Blot, much less a thermometer. It is unclear what the “enforceability” scheme measures, how well it measures it, or whether it is supposed to measure the same thing in each study that employing it.

The 2 studies are, in many ways, thoughtful attempts to investigate physician NCTs. But they address complex topics at the intersection of law, policy, and both labor and industrial organization economics. Recalling general concerns about the literature, both physician studies rely on survey data, both evidence selection and sampling issues, and both are subject to significant additional limitations. ⁷ Hausman and Lavetti depend critically on the confounding “enforceability” measurement scheme from Bishara; Lavetti et al do provide some results independent of “enforceability,” although their more striking result depends on it.

Finally, both staff and management in the FTC’s Bureau of Economics have made substantial contributions to the study of competition in health care markets, with special focus on the study of provider consolidation.^{39 -45} Yet Hausman and Lavetti employ both market definitions and analytic methods that are disfavored—or discredited—in contemporary antitrust analysis, including some specifically repudiated by research conducted by FTC staff. For example, their modeling approach and analysis depend on HHI measures of concentration and the structure-conduct-performance (SCP) framework that, based on both theoretical and empirical advances, have been largely abandoned by industrial organization economics and antitrust in recent decades.^{44,45,46 -51} Considerable research from the FTC’s Bureau of Economics has specifically undermined the SCP approach, even as a screening method, for provider mergers.^44,45,49 The point is not that concentration measures cannot have any signaling value. ⁵² Significant changes in concentration in markets that are, already, highly concentrated may raise competitive concerns. Indeed, it should be noted that new Merger Guidelines, published jointly by the FTC and the Department of Justice, suggest that presumptions of illegality should be attached to mergers that result in “highly concentrated” markets and involve a certain level of increase in HHI. ³⁷ Rather, it is that the general utility of the SCP paradigm and HHI measures, as applied in the research at issue, among other contexts, is in doubt.

It seems plain that further research is needed to warrant a general policy response to physician NCTs. In brief, general concerns about unintended consequences and error costs suggest, at least, that broad one-size-fits-all interventions in labor markets at a national level demand a more mature, and more settled, body of literature than is currently available. Such a normative demand militates against, for example, the FTC’s proposed NCT rule, which would apply nationally, across all occupations and income tranches. It also militates against a simple uniform rule applying to roughly 1 million U.S. physicians, given a very limited physician-specific empirical literature on the one hand and a highly varied services market on the other, with physicians providing a diverse range of services to patients in diverse market settings under varied business and employment models.

An Interim Approach

Research is not a policy solution, but it should be prolog to effective policy reform, and there is every reason to think that further study of physician NCTs will be useful. For one, ongoing state law reforms can be construed as “natural experiments”—policy reform events suitable for event studies that do not rely on a “black box” enforceability measurement tool. As noted above, some 15 states and the District of Columbia have statutory provisions bearing on physician NCTs, and recent statutory changes in, for example, Colorado and the District of Columbia present distinct events for study with established methods. Recent inclusion of a question regarding NCTs in the National Longitudinal Survey of Youth is a salutary addition to the survey evidence on NCT usage. ²⁷ Still, given the policy focus on NCTs at the federal and state levels, additional investment in refinement and validation of the (yet unvalidated) survey instrument could improve the quality of the survey evidence. Finally, the FTC’s investigative authority under section 6(b) of the FTC Act could be employed to gather direct evidence of NCT usage, diminishing or perhaps obviating reliance on survey evidence.

Antitrust enforcement—not broad FTC rulemaking, but case-specific litigation under the “rule of reason”—may provide a useful means of intervention, as least as an interim measure. First, as explained below, rule of reason inquiry is an established—indeed, dominant—means of scrutinizing commercial restraints under the U.S. antitrust laws. ⁵³ Second, private plaintiffs, as well as federal and state enforcers, may bring suit under the Sherman Act.^54,55 Third, such actions, while challenging, ³ may be most tractable where physician plaintiffs have suffered damages and the benefits of health care price and quality competition are most at risk. That is, one might demonstrate market-wide labor monopsony effects in highly concentrated physician services markets, which will tend to be defined both in terms of a product market (specific bundles of services) and in terms of a geographic market (the geographic scope within which, eg, certain anesthesiology or pediatric practices compete for patients or other payers),^56,57 although competitive effects may be shown directly. Further, one might do so based on market-specific evidence gathered in an enforcement investigation or, in the case of private enforcement, during discovery. Doing so would reduce reliance on the as-yet limited empirical literature and average effects suggested therein, and it would mitigate concerns about unintended consequences and error costs associated with general rulemaking. Litigation is costly, but private plaintiffs can recover treble damages in antitrust suits brought under the Sherman Act; and they may recover attorneys’ fees and, under certain limited circumstances, pre-judgment interest. ⁵⁴ Antitrust class-action suits can dramatically lower the costs imposed on individual plaintiffs; and successful ones can be highly effective in creating a credible threat of liability in other markets. The FTC may also contemplate action under its “standalone Section 5 authority.” ⁵⁸ It should be noted, however, that the FTC has commonly brought unilateral conduct cases alleging violations of both the Sherman Act and the FTC Act, and its recent Section 5 policy statement is both controversial and untested.^{59 -62}

Physicians may also be much better positioned than many other workers when it comes to recruiting assistance from state attorneys general (who can sue on behalf of their citizens in parens patriae) or federal enforcers. And federal enforcers might well be interested in physician NCT cases, given the agencies’ conspicuous new interest in labor antitrust matters and NCTs. The FTC, for example has brought (and settled) several complaints involving NCTs in other industries within the past year. ⁶³ Given the FTC’s broader health care competition enforcement program, it might bring or join physician NCT cases under the right facts and circumstances.

Most antitrust cases in the U.S. are decided under the “rule of reason,” ⁵³ first articulated by the Supreme Court in 1911, ⁶⁴ if subject to ongoing refinement.^52,65-67 In brief, the rule of reason requires that plaintiffs (including federal enforcers) plead and prove that defendants with market power have engaged in anticompetitive conduct; and anticompetitive conduct must have (or be likely to have) anticompetitive effects.^53,67 Typically, these are assessed in terms of higher prices or reduced output, although non-price factors such as reduced quality are also cognizable antitrust harms. The rule of reason is a sort of balancing of harms and benefits, although it does not entail a total welfare analysis or the pretense of a comprehensive cost-benefit analysis. If a plaintiff shows a likelihood of substantial cognizable harm—to “competition, not [just individual] competitors ⁶⁸ —the defendant has the burden of showing countervailing competitive benefits; doing so shifts the burden of proof back to the plaintiff. ⁶⁹

Two more qualifications are important. First, the requisite showing of consumer harm can be established by identifying immediate consumers or those downstream in the chain of production and distribution. Second, “consumers” can be suppliers in some contexts; that is, exploitation of monopsony power to harm sellers (say, of labor in a labor market) can also be construed as an antitrust violation under the rule of reason and the “consumer welfare” standard. ⁷⁰

The basic elements of 2 sorts of cases might be sketched in brief. First, agreements between competitors (“horizontal agreements”) may be subject to per se condemnation if they fix prices or output or divide markets. Hence, for example, the FTC and the Department of Justice have jointly issued guidance to Human Resources Professionals stating that naked wage-fixing or no-poaching agreements among competitors can be viewed as per se antitrust violations. ⁷¹ Similarly, an agreement between competing health care providers on the use of physician NCTs could be deemed a per se violation, to the extent it functions as a horizontal agreement to fix non-price terms of employment.

Second, agreements between institutional providers and physicians could be adjudged unlawful “vertical” agreements, to the extent that employers with market power use them to exclude competitor entry, and not merely as focused (appropriately tailored) means of protecting proprietary information or training investments. Consider, for example, a tertiary care hospital that is the only provider of cardiac surgery in a geographic market. For certain surgical procedures, the hospital is a monopolist; and market challenges to competitive entry can be compounded by regulatory barriers, such as state Certificate of Need Programs.^31,32

The simple fact of being a monopolist is not an antitrust violation. But suppose the monopolist requires its surgeons to sign NCTs, preventing them from leaving to establish a competing practice or, for example, staffing a would-be entrant into the market. Without pretending to conduct discovery and litigation over complex factual details, one can nonetheless see that the employer in such a case might use NCTs—successfully—to restrict competition, perhaps to prop up the prices it negotiates with health plans, suppress compensation for its physicians, or both. That might be actionable as unlawful exclusionary conduct.

Analysis of an antitrust claim under the rule of reason can be a complex, fact-specific task, and one subject to uncertainty before adjudication. The foregoing is not supposed to be a facile recipe for establishing that specific physician NCTs violate the antitrust laws. But there is a considerable body of economic literature regarding provider mergers and concentration, and a considerable body of case law on provider mergers and conduct, that can inform, if not determine, analysis of physician NCTs. This at least sketches the possibility of successful and procompetitive enforcement actions under circumstances where competition, consumers (patients), and physicians might benefit most; that is, where competition concerns for health care services markets are most pronounced and competitive concerns about those markets and physician labor markets are likely to co-vary.